The planned project will allow Geneva city to have a real-time dashboard with many parameters. The price-affordable pollution sensors have known issues, like quality decreasing with months, involving them to be replaced regularly. They also have error such as repetitive measurement which could decrease the precision. Also know the limitation of these sensors is the integration time to get the value ready which could be a few tens of seconds. In this case, the air sample should not be the brew for example due to moving vehicles.
This preliminary study may help respond to the first questions, which are, with a high quantity of data coming from the different sensor:
- Can we give a quality mark and trust these measurements with all these geo-localized data?
- Can we have compensated error precision with a high number of cross-checked data?
- Can algorithms increase sensor lifetime with a high number of trusted data?
This project could suggest an innovative methodology and algorithms that could be deployed at a large scale in a future project.